[
https://issues.apache.org/jira/browse/STORM-386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14120725#comment-14120725
]
ASF GitHub Bot commented on STORM-386:
--------------------------------------
Github user harshach commented on the pull request:
https://github.com/apache/incubator-storm/pull/177#issuecomment-54386752
@anyatch I don't see the above mentioned issue anymore. But for some reason
complete_latency on topology ui goes up very high
window emitted transferred complete latency(ms) acked failed
10m 0s 747380 453520 24592.787 801800 8600
concerned with high complete latency and failed tuples and after a while
topology stops
I understand that WordCountTopologyNode is an example but I think it should
be working properly as users can refer to it as documentation towards writing a
topology in node.js.
I'll try to find more details (new to node.js) this evening. If you can
check why there are failed tuples and complete latency is high that will be
great.
> Development of multilang protocol in nodejs
> -------------------------------------------
>
> Key: STORM-386
> URL: https://issues.apache.org/jira/browse/STORM-386
> Project: Apache Storm (Incubating)
> Issue Type: New Feature
> Environment: nodejs
> Reporter: Anya Tchernishov
>
> Support nodejs multilang protocol.
> Design considerations:
> - Emit will receive an object (like args and kwargs in python) and a callback
> that is called when task ids list is received.
> self.emit({tuple: [word]}, function(taskIds) {
> self.log(word + ' sent to task ids - ' + taskIds);
> });
> - The following methods will received done method that must be invoked on
> completion (same pattern used by the mocha test framework for async unit
> tests).
> - Internal implementation of emit uses Process.stdout.write without a
> callback since nodejs streams maintain FIFO order (so far, we have not found
> a need for providing a callback).
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
